The present invention pertains to new polynucleotides or new combinations of polynucleotides useful as diagnostic tools for predicting the occurrence of a human hepatocellular carcinoma disease. The invention is also directed to polynucleotides that consist in candidate tumor suppressor genes the alteration of which is involved in the occurrence of hepatocellar carcinoma in a patient, as well as to polynucletides derived from such new candidate tumor suppressor genes and to the corresponding expressed polypeptides. The invention also concerns diagnostic methods using said polynucleotides as diagnostic tools.
Hepatocellular carcinoma (HCC) is the most common primary liver cancer in the world, with 251,000 new cases each year (Bosh et al., 1991) and, to date, this pathology carries a very poor prognosis. Epidemiological evidence has shown the predominant role of hepatitis B virus (HBV) as a major causal agent of liver cancer. Other risk factors include chronic infection with hepatitis C virus (HCV), alcohol abuse, environmental exposure to hepatocarcinogens such as aflatoxin B1, and several genetic diseases (Reviewed in Buendia et al., 1995, and described also in Bosch et al., 1991; Wogan, 1992). More particularly, epidemiologic studies indicate that more than 50% of HCCs are attributable to chronic hepatitis B virus (HBV) infection (Bosch et al., 1991). However, the role of hepatotropic viral agents and the molecular events leading to liver carcinogenesis remain unknown. A mutagenic role of HBV DNA integration in the host genome, that occurs frequently at early stages of HBV-associated tumorigenesis, was conclusively established only in rare cases (Dejean et al., 1986; De Thxc3xa9 et al., 1987; Wang et al., 1990), suggesting more indirect transformation pathways (Matsubara, 1991). Viral DNA integrated into hepatocyte DNA can be detected in about 80% of chronic HBV carriers (Chen et al., 1986).
A common feature in chronic viral hepatitis and liver cirrhosis is long lasting inflammation of the liver associated with chronic regenerative conditions, which might enhance the susceptibility of liver cells to genetic changes. HCC usually develops after a 20-50 year period of HBV chronic infection, often subsequent to cirrhosis (Lok et al., 1991). The long latent period before the establishment of carcinomas indicates that they are the result of a multistep process, and several studies have been directed toward the identification of common genetic alterations (Sugimura et al., 1992). Both activation of cellular oncogenes and inactivation of tumor suppressor genes have been implicated (Okuda et al., 1992; Sugimura et al., 1992).
Generally, the development of human cancer results from clonal expansion of genetically modified cells that acquired selective growth advantage through accumulated alterations of ptoto-oncogenes and tumor suppressor genes (Weinberg, 1991). Somatic inactivation of tumor suppressor genes is usually achieved by intragenic mutations in one allele of the gene and by the loss of a chromosomal region spanning the second allele.
The steps that lead to homozygosity of a mutant suppressor allele usually involve the flanking chromosomal regions as well. Accordingly, anonymous DNA markers mapping to nearby chromosomal sites, which may have shown heterozygosity prior to tumor progression, will suffer a parallel reduction to homozygosity (or loss of heterozygosityxe2x80x94LOH). Indeed the repeated observation of LOH of a specific chromosomal marker in cells from a particular type suggests the presence of a closely mapping tumor suppressor gene, the loss of which is involved in tumor pathogenesis (Hansen et al., 1987). The recessive action of mutant suppressor gene alleles permits any resulting phenotypic effects to be delayed for long periods of time after conception. These alleles are effectively latent until they are exposed by a reduction to homozygosity in one or another cell.
Thus, a tumor suppressor gene is a genetic element whose loss or inactivation allows a cell to display one or another phenotype of neoplastic growth deregulation. Such a definition exclude genes that are cytostatic or cytotoxic when introduced into a cell and inappropriately overexpressed. The arena of action of tumor suppressor genes may thus be defined: biochemically, these genes serve as transducers of anti-proliferative signals; biologically, they serve as part of the response machinery that enables a cell to stop progression through the cell cycle, to differentiate, to senesce, or to die (Weinberg, 1991).
Chromosomal analysis using polymorphic DNA markers that distinguish different alleles has revealed loss of hereozygosity (LOH) of specific chromosomal regions in various types of cancers and the mapping of regions with a high frequency of LOH has been critical for identifying negative regulators of tumor growth (Call et al., 1990; Fearon et al., 1990; Friend et al., 1986). The recent development of microstaellite polymorphic markers has allowed positional cloning of several tumor suppressor genes such as the BRCA1, BRCA2 and DPC4 genes (Hahn et al., 1996; Miki et al., 1994; Wooster et al., 1995).
Previous studies, mainly relying upon either restriction fragment length polymorphism (RFLP) markers or microsatellites markers restricted to specific chromosome arms, have defined a number of chromosomal regions of LOH in liver cancer. One of the most frequent allelic deletions in HCC has been found at chromosome 17p where the tumor suppressor gene p53 is located (Fujimori et al., 1991; Murakami et al., 1991; Slagle et al., 1991). The frequency of p53 mutations varies largely among HCC samples, depending on the geographic location in the world, and a hot spot mutation at codon 249 was observed in HCCs from regions with high levels of dietary aflatoxins and high prevalence of HBV infection (Bressac et al., 1991; Buetow et al., 1992; Hsu et al., 1991). Regional deletions spanning the RB locus on chromosome 13q have also been described, but in this case, a low mutation rate was found in the remaining allele (Murakami et al., 1991; Wang and Rogler., 1988; Zhang et al., 1994). The most frequent chromosome arm deletion is observed in 13q (53% of informative tumors). Deletions were encompassing a large region of 13q (13q12-q32) which harbors the RB and BRCA2 tumor suppressor genes (Friend et al., 1986; Wooster et al., 1995; Zhang et al., 1994). Other frequent LOH was reported on chromosome arms 1p, 4q, 5q, 6q, 8p, 10q, 11p, 16p, 16q and 22q (Buetow et al., 1989; De Souza et al., 1995; Emi et al., 1992; Fujimori et al., 1991; Takahashi et al., 1993, Tsuda et al., 1990; Wang and Rogler, 1988; Yeh et al., 1994). Candidate tumor suppressor genes in these regions include the mannose 6-phosphate/insulin-like growth factor II receptor gene (M6P/IGF2R) on 6q26-q27 (De Souza et al., 1995), the PDGF-receptor beta-like tumor suppressor gene (PRLTS) on 8p21-p22 (Fujiwara et al., 1995) and the E-Cadherin gene on 16q22 (Slagle et al., 1993).
Yeh et al. (1994) have performed a genetic analysis of HCC cell lines and 30 primary HCC tissues. Using 8 Polymorphic DNA markers for RFLP experiments and also microstaellites markers spanning 12 loci in chromosome 1p, these authors have shown that main chromosomal abnormalities seemed to cluster at the distal part of chromosome 1p, with a common region mapped to 1p35-36, which is also the region with frequent loss of heterozygosity in neuroblastoma and colorectal as well as breast cancers.
Tsuda et al. (1990) have studied allele loss on chromosme 16 by performing RFLP analysis of 70 surgically resected tumors by using 15 polymorphic DNA markers distributed overall both the short arm and the long arm of said chromosome. They detected LOH in 52% of informative cases (i.e. 36 cases), the common region of allele loss being located between the HP locus (16q22.1) and the CTRB locus (16q22.3-q23.2).
Fujimori et al. (1991) have realized an allelotype study of HCC by examining LOH with 44 RFLP markers in 46 cases of HCC. The markers used by Fujimori et al; represented all chromosomal arms excepted 5p, 8p, 9p, 18p and acrocentric chromosomes. Each chromosomal arm was thus mapped with only a single or two polymorphic RFLP markers. These authors have observed that a significant percentage of LOH occurred for chromosome arms 5q (4 deletions in 9 informative cases [44% LOH]), 10q (6 deletions in 24 informative cases[25% LOH]), 11p (6 deletions in 13 informative cases [46% LOH]), 16q (12 deletions in 33 informative cases[36% LOH]) and 17p (5 deletions in 11 informative cases [45% LOH]).
Buetow et al., (Buetow et al., 1989) reported LOH at the albumin gene locus (4q11-q12) in all of five informative HCCs, indicating that a tumor suppressor gene might lie in this region. The inventor""s data suggest that alterations in two additional loci on chromosome 4q may play a role in liver carcinogenesis. Because chromosome 4q contains genes encoding growth factors or genes expressed predominantly in the liver such as albumin, alcohol dehydrogenase (ADH3), fibrinogen and UDP-glucuronyl-transferase, the deletion of this region might profoundly alter cell growth conditions and hepatocyte functions.
Buetow et al. (1989) have studied the LOH in 12 human primary liver tumors that have been tested against a panel of RFLP markers. These authors have typed tumor and non tumor issue for 11 RFLP markers spanning from 4q11-q13 to 4q32 chromosome 4 regions. In addition, Buetow et al. tested at least one RFLP marker on nine other chromosomes (1, 2, 6, 7, 9, 11, 13, 14 and 17) for allelic loss. The results showed that seven on nine tumors constitutionally heterozygous for chromosome 4q markers (six 4q RFLP markers were used by Buetow et al.) showed allele loss in tumor tissue. Six of the seven sample were jointly informative for both 4p and 4q markers (six 4p RFLP markers used). Among the other chromosomes informative for allele loss, one tumor showed changes in 13q. No other changes were observed in RFLP markers located on the eight other chromosomes tested. These authors concluded that a controlling locus involved in the pathogenesis of HCC might be in the vicinity of 4q32.
Emi et al. (1992) observed a frequent LOH for different loci on chromosome 8p in tumor tissues derived from HCC, colorectal cancer and lung cancer. More particularly, Emi et al. studied LOH in 120 HCC (46 of which had previously already been allelotyped by Fujimori et al. in 1991) tissues with five polymorphic markers along the short arm of chromosome 8 and defined commonly deleted regions within the same chromosomal interval, 8p23.1 to 8p21.3, suggesting that one or more tumor suppressor genes for HCC, and also for colorectal cancer, might be present in said region. The region of interest was mapped by Emi et al. using only three RFLP polymorphic DNA markers, respectively D8S238, MSR and D8S220. These authors concluded that a putative tumor suppressor gene might exist on 8p.
Becker et al. (1996), in order to investigate the chromosome 8 allele status in Chinese HCC, described that a panel of 37 matched normal and HCC DNAs from Qidong was analyzed for tumor specific allele loss with eight specific RFLP probes from both arms of chromosome 8. Tumor-specific LOH was found highest on the short arm with 71.4% (10/14) and 85% (17/20) of the informative patients missing an allele for 8p23 or 8p21 (only two RFLP specific probes used for the entire chromosome 8 short arm), respectively. Allele loss from the long arm of chromosome 8 was also observed with 30% (6/20) and 33.3% (7/21) of the samples informative for 8q22 and 8q24, respectively.
Boige et al., in 1996, have studied the allelic deletions in HCC, using 275 higly polymorphic microsatellites genetic markers spanning all non acrocentric chromosome arms in a group of 48 HCC. They observed that nine chromosome arms were deleted in more than 30% in 1p, 1q, 4q, 6q, 8p, 9p, 16p, 16q and 17p, the most frequent chromosome arm deletion being observed for 8p.
The scientific works described hereinbefore have allowed a primary but very coarse localization of the different genetic alterations occurring on the different chromosome arms present in the HCC tissue samples, due either to the weak number of genetic markers used and to the weak number of patients"" tissue samples studied. The weak number of patients"" tissue sample used in these studies did not provide conclusive or statistically significant data as to the frequence of a genetic alteration on a given chromosome arm or chromosome region. The poor precision with which the altered focuses were identified, and consequently the great size of the chromosome DNA fragments of interest that are bordered by the polymorphic DNA markers used in these studies, did not allow the one ordinary skill in the art to design and/or determine the suitable technical means useful to design accurate diagnostic tools for HCC, because none of the DNA fragments was shown in prior art to be sufficiently relevant and be considered as carrying the causal information for operating a precise correlation with the disease. Consequently, they did not permit the one ordinary skill in the art to clone specific DNA fragments from the chromosomes of healthy tissues, that were observed to be frequently altered during the occurrence of HCC.
Primary liver tumors, like other solid tumors in humans, most likely arise through a cascade of genetic events involving oncogenes and tumor suppressor genes that results in decreasing stability of the genome and ultimately leads to the malignant phenotype. The methods used according to the invention, that allowed an accurate and complete scan of 120 hepatocellular carcinoma genomes for allelic imbalance, are of decisive value for locating candidate genes implicated in liver cancer development.
The inventors have detected frequent changes in allelic dosage at particular regions of chromosomes 1p, 4q, 6q, 8p, 13q, 16p, 16q and 17p. In addition, the invention describes newly identified loci exhibiting a high rate of LOH on chromosome arms 1q, 2q, 9p, 9q, 14q and 17q. Genetic alterations of these regions have been observed in many other malignancies including breast cancers, small-cell lung carcinomas, prostate cancers, renal cell carcinomas, malignant melanomas or meningiomas, suggesting that tumor suppressor genes implicated in a wide spectrum of tumors may affect HCC progression (Cox et al., 1995; Hearly et al., 1995; Ranford et al., 1995; Takahashi et al., 1995; Takita et al., 1995).
The inventor""s data indicate that a more telomeric locus than 8p21-p23, which shows a higher frequency of LOH compared to 8p21-p23 markers, is likely to contain a second candidate tumor suppressor gene for HCC. On the long arm of chromosome 4, the inventors have defined three non contiguous regions of LOH in 4q12, 4q22-q24 and 4q35. Chromosomal alterations at 4q are of great diagnostic value because they represent a preferential trait in HCC.
Further, the inventors have found frequent LOH with markers spanning a large region of 6q (Table 3 and data not shown).
The inventors have also found frequent LOH with markers depicted in Tables 2 and 4, therefore allowing the one skill in the art to determine the presence and characteristics of HCC-associated tumor suppressor genes existing on the different identified loci of interest, particularly on 8p21-8p23, 1p35-p36, 16q23-q24, 14q32, 4q and 6q.
Correlation of LOH with histological analysis of the adjacent non tumorous livers revealed that the frequency of deletions at combined chromosomal regions (1p, 6q, 8p and 13q) is significantly higher in HCCs arising on chronic hepatitis lesions than in HCCs developing on liver cirrhosis. It is here shown that when hepatocytes accumulate allelic changes under regenerative pressure, some combined chromosomal arm losses might constitute a xe2x80x9cshort-cutxe2x80x9d towards malignancy.
The present inventors have now used a very dense panel of polymorphic microsatellite markers in order to precisely localize the different chromosomes regions being altered in HCC patients cancer liver tissue samples. The inventors have therefore performed a systematic screening of 120 HCC samples using 256 highly polymorphic microsatellite markers evenly distributed throughout non-acrocentric human autosomes. This is the first, large scale analysis of genomic alterations in human HCC using microsatellite markers. The data allow precise estimation of the relative frequency and accurate positioning of each chromosomal change.
The present invention thus concerns the use of DNA markers localized in specific chromosomal loci of interest (that are defined hereunder) as diagnostic tools allowing the prognosis of the hepatocellular carcinoma (HCC).
Because the inventors have now precisely identified the very small chromosomal regions that are subject to frequent genetic alterations, the DNA markers embraced by the present invention cover all the publicly available tools spanning these specific chromosomal loci of interest, namely:
1) Microsatellite DNA markers;
2) RFLP markers (usually constituted by specific oligonucleotide probes);
3) VNTR markers (Variable Number of Tandem Repeats), also named  less than  less than ministallites greater than  greater than  that are sequences of the  less than  less than Alu greater than  greater than  type of about twenty nucleotides that are repeated at high number of copies inside each VNTR, and which are detected either by a PCR reaction or a Southern blot hybridization;
4) STSs markers (Simple Tag Sequences), which are unique genomic sequences that can be amplified by a pair of specific oligonucleotide primers and which are generally non polymorphic;
5) ESTs (Expressed sequence Tags) which are transcribed in mRNA and that can be amplified by a pair of oligonucleotide primers.
The sequences of the DNA markers of the above groups 1) to 4) as well as the sequences of the oligonucleotide detection tools for each of them are freely publicly available on electronic databases, particularly on the Internet World Wide Web at the following address:  less than  less than http://www.ncbi.nlm.nih.gov) greater than  greater than .
The present invention also concerns the use of polymorphic microsatellite DNA markers as diagnostic tools allowing the prognosis of the hepatocellular carcinoma (HCC).
The present invention is also directed to diagnostic methods using such microsatellite DNA markers, as well as to diagnostic kits comprising these polymorphic DNA markers and the reagents necessary to perform the diagnostic methods of the invention.